VGG16 Feature exctraction and training on medica dataset

feature extraction plus ml algorithm.py

# -*- coding: utf-8 -*-

''' Importar bibliotecas'''
import keras
from keras.preprocessing import image
from keras.applications.vgg16 import VGG16
from keras.applications.vgg16 import preprocess_input
from sklearn.model_selection import train_test_split
import numpy as np
from keras.models import Sequential
from keras.layers.core import Dense,Dropout
import matplotlib.pyplot as plt
#import datetime

# 0,1,2,3
classes = ["blurry-nothing","colon-clear","polyps","esophagitis"]
path = "Medico_2018_development_set/"
#images_per_class = 150
images_per_class = 150
#path puede ser Medico_2018_development_set/+class for class in classes

model = VGG16(include_top=False)
#model.summary()



def extract_features():
    vgg16_feature_list = []
    labels = []
    j = 0
    for class_ in classes:
        #vgg16_feature_list[str(class_)] = []
        #Tomar el timepo por clase
        #start_time = datetime.datetime.now()
        for i in range(1,images_per_class+1):
            current_path = path+class_+"/"+str(i)+".jpg"
            img= image.load_img(current_path,target_size=(224,224))
            img_data = image.img_to_array(img)
            img_data = np.expand_dims(img_data,axis=0)
            img_data = preprocess_input(img_data)
            ith_feature = np.array(model.predict(img_data))
            #ith_feature_np = np.array(ith_feature)
            vgg16_feature_list.append(ith_feature.flatten())
            labels.append(j)
        j = j+1
    vgg16_feature_list_np = np.array(vgg16_feature_list)
    labels_np = np.array(labels)
    return (vgg16_feature_list_np,labels_np)
        #features[str(class_)] = vgg16_feature_list_np
        #end_time = datetime.datetime.now()
        #elapsed = end_time - start_time
        #print("tiempo para clase "+str(class_))
        #short_microseconds = elapsed.microseconds/(10**4)
    #print("%d:%d s"%(elapsed.seconds,short_microseconds)) 
    #print("-"*10)
#vgg16_feature_list_np = np.array(vgg16_feature_list)
''' Final feature list
features
'''
dataset,labels = extract_features()
''' 
print(dataset.shape) (40, 25088)
print(labels.shape) (40,)
'''

#--------PREPARAR DATOS-----------------
x_train,x_test,y_train,y_test = train_test_split(dataset,labels,test_size=0.25,random_state=4)
num_classes = len(classes)
#One-hot encoding
y_train = keras.utils.to_categorical(y_train, num_classes)
y_test = keras.utils.to_categorical(y_test, num_classes)


#----------ML ALGO-----
input_shape = (25088,)

model = Sequential()
model.add(Dense(256, activation='relu', input_shape=input_shape))
model.add(Dropout(0.5))
model.add(Dense(num_classes, activation='softmax'))
model.compile(loss=keras.losses.categorical_crossentropy,
              optimizer=keras.optimizers.Adadelta(),
              metrics=['accuracy'])

history = model.fit(x_train,y_train,epochs=20,validation_data=(x_test,y_test))




acc = history.history['acc']
val_acc = history.history['val_acc']
loss = history.history['loss']
val_loss = history.history['val_loss']

epochs = range(len(acc))

plt.plot(epochs, acc, 'bo', label='Training acc')
plt.plot(epochs, val_acc, 'b', label='Validation acc')
plt.title('Training and validation accuracy')
plt.legend()

plt.figure()

plt.plot(epochs, loss, 'bo', label='Training loss')
plt.plot(epochs, val_loss, 'b', label='Validation loss')
plt.title('Training and validation loss')
plt.legend()

plt.show()
score = model.evaluate(x_test, y_test, verbose=0)
print('Test loss:', score[0])
print('Test accuracy:', score[1])